To date, there have been few studies involving the efficacy of combinations of antiretrovirals in HIV-infected children who are antiretroviral experienced and stable. And currently, the duration of efficacy of continuing stable therapy in children is undefined. In addition, the indications for timing of changes of antiretroviral therapy in children are not clear. It is therefore important for the durability and efficacy of new therapeutic combinations of antiretrovirals be identified for children. The goal of this trial is to compare the efficacy of 4 different combinations of antiretroviral agents to reduce viral replication by 2.2 logs or more at 12 weeks and the ability of the combination to maintain viral suppression to "undetectable" levels in children who are clinically stable and antiretroviral experienced. The 4 combinations to be tested in this multi- center trial involving 240 clinically stable, HIV-infected children between the ages of 4 months and 17 years include: Arm A: d4T, nevirapine, ritonavir Arm B: d4T, 3TC, nefinavir Arm C: d4T, nevirapine, nelfinavir Arm D: d4T, 3TC, nevirapine, nelfinavir One of the earliest defects in immune function after HIV infection is a loss of cell-mediated responses to recall antigens, even when CD4 attrition is not evident. In addition, there appears to be a progressive loss of lymphoproliferative responses to HIV antigens with advancing Walter Reed stages. Until recently, successful responses to therapeutic interventions have been assessed by measuring incremental or decremental changes in CD4 T cell numbers or percentages. It was never clear whether any increase in CD4 T cells resulted in improved function, particularly since it is unknown if the recovering cells are immunologically naive or experienced. A recent study noted some reconstitution of helper T cell function after ZDV therapy, without correlation to improved CD4 T cell numbers. Since the ultimate goal of Antiretroviral therapy is to normalize immune function, a measurement of T cell function would be a valid criteria for determining treatment efficacy. Studies of lymphocyte subsets in HIV infected children suggest that naive CD45RA subsets of both CD4 (Helper) and CD8 (cytotoxic suppressor) T cells are depleted with relative preservation of CD45RO (Memory) T cells. This observation may be related to decreased ability to respond to new antigenic stimuli. In adults treated with ritonavir, Kelleher noted significant increases in naive and memory CD4 cells, and improved lymphocyte proliferation responses to mitogen, recall antigens and HIV-specific proteins. In addition several preliminary studies have suggested significant changes in expression of activation markers such as CD38 and DR as well as altered expression of CD28 and up regulation of L selectin CD621 on the lymphocytes of HIV infected adults. Expression of these markers in vitro may act as surrogate indicators of disease progression and immune reconstitution in subjects treated with HAART regimens. The effects of these therapies on T cell numbers and function have not been studied in large numbers of children. Preliminary data on the effects of combination therapy with NRTIs and Protease inhibitors should be forthcoming from PRAM-1 (ACTG 338). However, study of these surface markers and in vitro LPR in the highly active regimens contained in this protocol will add substantial information on regimens designed to rapidly reduce viral loads to undetectable levels by addition of protease inhibitors to the reverse transcriptase inhibitors.